The present invention relates to a method and a system for collecting the performance data of hardware devices constituting a storage network and the software operated in the hardware devices, or in particular a method and a system for collecting the storage network performance data suitable to a case in which the network is increased in scale to such an extent that the component elements for which the performance data are to be collected are vast in number.
A storage network so configured that a centralized storage is accessed from a plurality of host servers through the network is extending widely as an architecture for a data center to improve the utilization rate and reduce the management cost of the storage ever on the increase in scale.
The performance management software meets this situation by being configured of an agent arranged in a network for each hardware device or software for which the performance is to be monitored, and the management software for centrally managing the performance data for the whole network. Each agent acquires the performance data by direct communication with each object to be monitored, while the management software collects and accumulates the performance data acquired by the agents and supplies the performance data in response to a request of a storage network manager or the like.
Apart from the storage network, take a computer network as an example. A method and a system having a similar configuration to the above-mentioned method and system for monitoring the performance of a plurality of server devices in a network environment are disclosed in U.S. Pat. No. 6,505,248.
With the extension of the centralized storage based on a storage network, the component elements of the network increased in scale has become vast in number and the correlation between the component elements tends to be complicated more and more.
In order to monitor the performance of an application system and carries out the tuning in this storage network environment, the performance data for various hardware devices and software making up the network are required to be comprehensively collected and the correlation between them and the temporal change thereof are required to be grasped.
A technique for automating the collection of the dispersed performance data is indispensable for the performance management of this kind of the storage network. With a further increase expected in the scale of the network, however, automatic comprehensive collection of the performance data for all the component elements of the network may become considerably difficult in terms of the processing capacity including the storage capacity, computation performance and the communication performance.
In order to monitor and tune the performance of an application system in a large storage network environment, it is necessary to collect the performance data on the various hardware devices and software making up the network comprehensively and to grasp the correlation between them and the temporal change thereof.
This is by reason of the fact that unlike in the conventional architecture in which each application system is independently associated with a corresponding server with a computer processing system and an external storage connected directly to each other, the storage network environment is liable to develop an interference in performance between application systems at a portion shared by the network devices and the storage systems.
In some conventional techniques, the collecting operation for the performance data can be switched on/off for each network component element by manual updating operation of the user. The use of this function could limit the amount of the performance data to be collected. For this purpose, however, elements to be emphasized and elements to be ignored are required to be discriminated from each other in advance.
This is a considerably tough job for a storage network environment in which various applications having different tendencies, of the performance load are unified and a vast number of component elements affect each other in complicated way. Also, the manual operation of the user may cause the timing of acquiring crucial information to be lost or a problem, if any, to be detected too late.